g.stats.glob: Likelihood ratio G-statistic over loci

In hierfstat: Estimation and Tests of Hierarchical F-Statistics

Likelihood ratio G-statistic over loci

Description

Calculates the likelihood ratio G-statistic on a contingency table of alleles at one locus X sampling unit, and sums this statistic over the loci provided. The sampling unit could be any hierarchical level (patch, locality, region,...). By default, diploid data are assumed

Usage

g.stats.glob(data,diploid=TRUE)

Arguments

data	a data frame made of nl+1 column, nl being the number of loci. The first column contains the sampling unit, the others the multi-locus genotype. Only complete multi-locus genotypes are kept for calculation
diploid	Whether the data are from diploid (default) organisms

Value

g.stats.l	Per locus likelihood ratio statistic
g.stats	Overall loci likelihood ratio statistic

Author(s)

Jerome Goudet, DEE, UNIL, CH-1015 Lausanne Switzerland

[e-mail]

References

Goudet J. (2005). Hierfstat, a package for R to compute and test variance components and F-statistics. Molecular Ecology Notes. 5:184-186

Goudet J., Raymond, M., DeMeeus, T. and Rousset F. (1996) Testing differentiation in diploid populations. Genetics. 144: 1933-1940

Petit E., Balloux F. and Goudet J.(2001) Sex-biased dispersal in a migratory bat: A characterization using sex-specific demographic parameters. Evolution 55: 635-640.

See Also

g.stats, samp.within,samp.between.

Examples

## Not run: 
data(gtrunchier)
attach(gtrunchier)
nperm<-99
nobs<-length(Patch)
gglobs.o<-vector(length=(nperm+1))
gglobs.p<-vector(length=(nperm+1))
gglobs.l<-vector(length=(nperm+1))

gglobs.o[nperm+1]<-g.stats.glob(data.frame(Patch,gtrunchier[,-c(1,2)]))$g.stats
gglobs.p[nperm+1]<-g.stats.glob(data.frame(Patch,gtrunchier[,-c(1,2)]))$g.stats
gglobs.l[nperm+1]<-g.stats.glob(data.frame(Locality,gtrunchier[,-c(1,2)]))$g.stats

for (i in 1:nperm) #careful, might take a while
{
  gglobs.o[i]<-g.stats.glob(data.frame(Patch,gtrunchier[sample(Patch),-c(1,2)]))$g.stats
  gglobs.p[i]<-g.stats.glob(data.frame(Patch,gtrunchier[samp.within(Locality),-c(1,2)]))$g.stats
  gglobs.l[i]<-g.stats.glob(data.frame(Locality,gtrunchier[samp.between(Patch),-c(1,2)]))$g.stats
}
#p-value of first test (among patches)
p.globs.o<-sum(gglobs.o>=gglobs.o[nperm+1])/(nperm+1) 

#p-value of second test (among patches within localities)
p.globs.p<-sum(gglobs.p>=gglobs.p[nperm+1])/(nperm+1) 

#p-value of third test (among localities)
p.globs.l<-sum(gglobs.l>=gglobs.l[nperm+1])/(nperm+1) 


#Are alleles associated at random among patches
p.globs.o 

#Are alleles associated at random among patches within localities?
#Tests differentiation among patches within localities
p.globs.p 

#Are alleles associated at random among localities, keeping patches as one unit?
#Tests differentiation among localities
p.globs.l 

## End(Not run)

hierfstat documentation built on May 6, 2022, 1:05 a.m.